sci.physics.electromag

On Feb 29, 3:32 pm, Uncle Al wrote:
> Benj wrote:

> > 1. What is the magnetic field right at the surface of the magnet (with
> > the magnet in free space)
>
> Depends on bulk magentization and radius of curvature of the surface,
> hence pole pieces. A field has amplitude, gradient, divergence, and
> curl. Which of those are interesting to you?

Wow! Uncle Al being reasonable for a change! I've got to use this
opportunity!

I am interested in amplitude and gradient.

> A near-field dipole drops off as r^3. Don't bring two strong magnets'
> opposite poles into proximity. Very, very, very bad idea.

Actually the magnet near the faces (or pole pieces if used) isn't a
dipole. One actually has to get to a distance where magnet dimensions
are small compared to the distance. I'm interested in trying to
create a region of reasonably strong but uniform field with permanent
magnets instead of wire, electric current and heat. And yes, the idea
is to bring two opposite poles into proximity (hopefully with a non-
magnetic spacer present that is NOT a finger or other protuberance. )

> Depends. If you want a clean volume of uniform intense field, go for
> a Helmholtz pair, cheaply with ferrites. Get a dead microwave oven
> and recover the gyrotron.

Yes this is it. I want a Helmholtz idea but with strong field and no
heavy current. Is there a permanent magnet arrangement called a
"Helmholtz pair"? It seems to me that the first cut would be two
cylindrical permanent magnets (especially the strong ones from United
Nuclear) with a space between them. That then brings up the next
stage which is how does one "flatten" the field in the center region.
Absolute field strength is not as important as flux lines being
straight lines through the region. Since the field will fall off from
the centerline, I presume that if one shapes the magnets (or pole
pieces) to have less space on the edges of the cylinder than the
center you can compensate for the drop-off with the right shape.

> If you want maximum sturm und drang, a linear or shaped Halbach
> array. A Tesla in a one-inch bore will set you back $300,

Neat thing, but $300! OUCH! Anyway not interested in magnetic
bearings.

>
>
> Ignore the social bullshit

> Lifting force is field divergence. If you want big lift at a given
> field you need tiny radii of curvature - a ribbed surface.

Nah. Care about field strength not lifting force. I was only bringing
that up because that was the only parameter the seller listed on the
various magnets. I hoped that it might be related to B at the poles in
some manner.

> Why a rod? A donut makes more sense. You can buy switchable ceramic
> lifting magnet arrays from surplus places, or pay double at Edmund.

A rod because that is one of the shapes they sell. Also I figure it
gives a bit more field strength at the center then a donut. However,
you are right that a donut might with the best spacing be a sort of
"compensator" for the falling field from the centerline. However
shaped magnets or pole pieces probably gives more precise control.

What I like about the neodymium magnets is the high fields. I don't
think ordinary ceramic magnets are as strong. I've got some of those
already and wasn't too impressed. Yeah you can get them lots of
places including old microwaves and large speakers.

So as I see where I am, from the online calculator it appears easy to
get several thousand gauss over a reasonable volume. By shaping the
magnets (or adding shaped pole pieces) it should be possible to
compensate for field fall-off away from centerline. The field varies
along the axial direction, of course, but that is not a huge problem
so long as it doesn't fall below a high value. The direction of the
flux lines (ions spiral around them) is the important thing. The big
puzzle here is the actual calculation that goes between magnet shape
and the external field. And the "final solution" would be a
calculation to determine what shape magnet pair would be optimum to
emulate the uniform field characteristics of a pair of Helmholtz
coils. Sounds like a patent waiting to be applied for... I guess
computers are good for something!

Benj